Reading Time: 3 minutes

We conclude the series of articles on how the AI revolution has modified and continues to modify the IT Talent environment, today moving from the realm of professional talent to the realm of companies.

Just as important as a certified professional who conveys the necessary confidence in their skills, is to be able to be sure that the organization or company developing products and services associated with AI can demonstrate a reasonable level of auditing and certification.

In this environment, the investment and return are not yet so clear, especially because many of the certifications or audits that are launched do not yet have the certainty of their continuity in the medium term. In the changing environment and exponential development of AI it is quite logical that this is the case.

A good basis to start with is to be already certified or “have the seal”, as they say, of standards already well tested and widespread in information security, process management and quality (ISO 27001, ISO 9001, National Security Scheme, etc.). If our organization has already passed through here, facing the more IA and Data oriented certifications will be much simpler, and above all much more accessible.

Let’s not forget that some of the aspects that most concern large companies when adopting AI, in addition to the ethical implications, are those aspects related to compliance, data security and location, privacy, etc.

Standards and Certification Organizations

ISO (International Organization for Standardization)

• ISO/IEC JTC 1/SC 42 Certification

• Description: This ISO committee develops international standards for artificial intelligence, including aspects of security, reliability and ethics.

IEEE (Institute of Electrical and Electronics Engineers)

• IEEE Certified AI Professional Certification

• Description: IEEE offers a number of courses and certifications focused on artificial intelligence and AI ethics.

We focus on the ISO because of its international diffusion mainly.

ISO/IEC JTC 1/SC 42 is a joint subcommittee of the International Organization for Standardization (ISO) and the International Electrotechnical Commission (IEC) dedicated to standardization in the field of artificial intelligence. More details on its functions, focus areas and the standards it develops are presented below:

ISO/IEC JTC 1/SC 42: General Overview

Functions and Objectives

• Establishment of Standards: Develop international standards that address artificial intelligence technologies and applications.
• Coordination: Coordinate with other ISO and IEC technical committees and subcommittees, as well as other organizations, to ensure consistency and avoid duplication.
• Evaluation and Audit: Evaluate the social, legal and ethical implications of artificial intelligence technologies.
• Adoption Facilitation: Facilitate the adoption of AI standards by industry, governments and other bodies.

Areas of Focus

• Big Data: Standards related to the management and analysis of large volumes of data.
• Machine Learning: Standards for the development, training, evaluation and application of machine learning models.
• Governance and Ethics: Guidelines and standards on the responsible and ethical use of AI.
• Trusted AI: Standards that ensure transparency, explainability, security and privacy in AI systems.

Main Standards Developed

• ISO/IEC 22989:2022 – AI Concepts and Terminology: Establishes common terminology and fundamental concepts in the field of artificial intelligence, providing a uniform basis for the development of other AI standards.
• ISO/IEC 23053:2022 – AI Framework: Provides a general framework for the development and implementation of artificial intelligence systems, covering aspects such as architecture, life cycle and best practices.
• ISO/IEC 24027:2020 – Data Quality Assessment for Machine Learning: Standards for assessing the quality of data used in the training and validation of machine learning models.
• ISO/IEC 20546:2019 – Big Data Overview and Vocabulary: Provides an overview and standard vocabulary for key terms and concepts related to Big Data.
• ISO/IEC TR 24028:2020 – Assessment of Machine Learning Classification Performance: Guidelines for the assessment of the performance of machine learning classification models, including metrics and evaluation methods.
• ISO/IEC TR 24030:2021 – Implementation of AI: Provides guidelines for the implementation of AI systems, covering technical, organizational and ethical aspects.
• ISO/IEC TR 24028:2021 – Guidelines on AI Ethical and Societal Considerations: Guidelines on ethical and societal considerations in AI development and implementation, including issues such as transparency, accountability and inclusiveness.

Relevance and Application

• Industry: The standards developed by ISO/IEC JTC 1/SC 42 are crucial for industry, as they provide a structured and recognized framework for developing and evaluating AI technologies.
• Governments: Governments can use these standards to formulate policies and regulations to ensure the ethical and responsible development of AI.
• Academia and Research: Academic and research institutions can adopt these standards to guide their projects and ensure interoperability and ethics in their work.
• Society: By addressing ethical and governance issues, these standards help mitigate risks and ensure that AI technologies benefit society as a whole.

Participation and Continuous Development

The ISO/IEC JTC 1/SC 42 subcommittee works continuously to develop new standards and update existing ones, based on evolving technology and market needs. Members include experts from various countries and organizations, ensuring a global and multidisciplinary representation in the standardization process.

In this context, with the speed of development of AI and Data Science and their interrelationship, standards already in use continue to be refined and new ones developed to meet the challenges that AI is generating. 

When it comes to overcoming possible barriers to AI adoption, the certification and auditing framework is a fundamental element, and it is also the one that is developing more slowly, but it is also advancing continuously to provide a secure framework for AI application and to avoid problems that may undoubtedly arise in the future.